AI ECONOMIC SALVATION OR SOCIETAL PERIL? Talk_April 2018.pdf · Accenture Research Bath Royal...

AI – ECONOMIC SALVATION OR

SOCIETAL PERIL?Should we welcome or fear the arrival of artificial intelligence?

Mark PurdyManaging Director, Economic Research

Accenture Research

Bath Royal Literary and Scientific Institution

23 April 2018Image: Kai Stachowiak

https://creativecommons.org/publicdomain/zero/1.0/

Copyright © 2018 Accenture. All rights reserved. 2

INTRODUCTION – SPEAKER

Mark Purdy,

Managing Director - Economic Research

Accenture Research


FOR MANY OF US, OUR FIRST INTRODUCTION TO THE IDEA

OF ARTIFICIAL INTELLIGENCE CAME VIA THE MOVIES

AI in service to the society or AI posing threats to humanity?

TARS, the space robot from ‘Interstellar’ (2014)

Robotics fighting back from ‘I, Robot’ (2002)

Crime forecasting from ‘Minority Report’ (2002)

Baymax, the healthcare robot from ‘Big Hero 6’ (2014)

A supercomputer deeming mankind a threat, Singularity (2017)

Machines taking control of Earth, ‘the Matrix’ (1999)

v.s.


HUMAN-LEVEL AI IS STILL A LONG WAY OFF

ARTIFICIAL GENERAL INTELLIGENCEARTIFICIAL NARROW INTELLIGENCE

Understand abstract concepts

Explain why

Be creative like children

Tell right from wrong

Have emotions

Beat Go world champions

Read facial expressions

Write music

Diagnose mental disorders

Comfort earthquake survivors

MORAVEC’S PARADOX

What AI can and can’t do


WHAT IS ARTIFICIAL INTELLIGENCE

Act

Sense

Comprehend

Computer Vision

Audio Processing

Natural Language Processing

Knowledge Representation

Machine Learning

Expert Systems

AI TECHNOLOGIES ILLUSTRATIVE SOLUTIONS

Virtual Agents

Identity Analytics

Cognitive Robotics

Speech Analytics

Recommendation

Systems

Data Visualization


AI HAS A LONG HISTORY OF SUMMERS AND WINTERS

19901950 1960 1970 1980 2000 2010

The 1st boom

• Reasoning as research

• Natural language

• WABOT-1-world’s first full-scale

intelligent humanoid robot

The birth of AI

• Cybernetics - thinking machines

• The Turning test

• Dartmouth conference

• Symbolic reasoning

The 1st AI winter

• Limited computer power

• Limited database capacity

• Limited networking capabilities

• Real-world problems are complicated

The 2nd boom

• The rise of expert systems

• The knowledge revolution

• Neural networks make a

comeback

The 2nd AI winter

• A sudden collapse of specialized AI

hardware vendors

• Nouvelle AI and embodied reason

2000-present

• Deep learning, big data and

artificial general intelligence


RECENTLY IT HAS BEEN UNDERGOING ANOTHER BOOM

AI Startup investment growth is outpacing that of

the overall startup market (2011-2016)

Global AI vs overall startup investment markets ($)

Overall market startup

investmentAI startup investment

54% 16%vs.

AI investments ($)

9X 6X

AI deals (#)

By 2016, 11% of all global startup deals were AI-oriented, up from 2% in 2010

0

2,000

4,000

6,000

8,000

10,000

Priority Year

2007

Num

be

r o

f P

rio

rity

Pa

ten

ts

2006 201020092008 2011 2012 2013 2014 2015c*

+24%

Filed Priority Patents – Artificial Intelligence

named filings and related technology filings

AI keywords (atleast one AI technology directly named in patent filing)

AI smartsearch (AI related tech filings)

*Note: For year 2015, only 6 months data is published and full year data is assumed at a

factor of two

Both AI startup investment and patent publication on AI have accelerated since 2010


BECAUSE OF THREE DRIVING FORCES

Rapid increases in

computational powerExplosion in data that is

digitized

Huge fall in cost of storing data –

near zero marginal cost

One dollar’s worth of computer power, 1980-2010

Billions of computations per second

(log scale)

Source: The Brookings Institution

Hard drive cost per gigabyte, 1980-2015

Global installed base of IoT units (number of units, billions)

0

5

10

15

20

25

2015

2016

2017f

2018f

2019f

2020f

Business Consumer

Source: Gartner, Forecast: Internet of Things - Endpoints and Associated Services, Worldwide, 2016

Source: mkomo.com

1,000,000

0.01

0.10

1,000

100,000

10,000

100

10

1

USD (log scale)

1980 1985 1990 1995 2000 2005 2010 2015


How people feel about AI?

• Public discussion has exploded since 2009

• In general, AI has had consistently more optimistic than

pessimistic coverage over time

THE RAPID ADVANCE OF AI HAS LED TO HEATED DEBATES

AROUND ITS POTENTIAL THREATS AND OPPORTUNITIES

Optimistic

articles

Pessimistic

articles

Articles on AI

Source: Long-term trends in the public perception of AI, Fast & Horvitz, 2016, Association for

the Advancement of Artificial Intelligence

AI in the news: The potential economic benefits and risks

10

COULD AI BE AN ECONOMIC

SALVATION?


GLOBAL ECONOMIC GROWTH IS WEAKENING

-6

-4

-2

0

2

4

6

8

10

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

Developed economies Emerging economies World

GDP growth (real, % y/y)

Source: Oxford Economics


LOOKING BACK, GDP GROWTH HAS STEADILY SLOWED IN

MANY LARGE ECONOMIES SINCE THE 1980S

1980s 1990s 2000s 2010s

Japan United States United Kingdom France Italy Germany

3.0

2.1

1.1 1.1

NB: Data points across the dashed lines indicate the average for the six countries.

Source: Oxford Economics

Real GDP growth (%, annual average over the period)


THIS IS BECAUSE THE TRADITIONAL FACTORS OF

PRODUCTION ARE UNDER PRESSURE

Capital efficiency: The marginal capital efficiency rate has steadily dropped over a 50-year period.

Marginal capital efficiency (%, 6-year moving average)

0

5

10

15

20

25

30

35

40

1961 1966 1971 1976 1981 1986 1991 1996 2001 2006 2011 2016

Labour:As populations age and birth rates slow, fewer people are available to pick up the slack in the workforce.

Working age population

(%, annual average growth over the period)

0.9

0.6

0.4

0.1

-0.1

-0.69

0.30.2

-0.05

-0.3

-0.7-0.73

UnitedStates

UnitedKingdom

France Italy Germany Japan

2000-2015 2016-2030

Japan United States Germany United Kingdom


…AND PRODUCTIVITY GROWTH IS TRENDING DOWNWARD

Total Factor Productivity (%, annual average over the period)

2.1

1.00.6

-0.1

0.7

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

1990-19941995-19992000-20042005-20092010-2014

0.4

-0.2

0.7

-0.9

-0.2

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

1990-19941995-19992000-20042005-20092010-2014

00.5 0.5

-0.7

0.1

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

1990-19941995-19992000-20042005-20092010-2014

Source: The Conference Board, Total Economic Database

0.5 0.71.1

0.040.5

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

1990-19941995-19992000-20042005-20092010-2014

0.3 0.1

0.9

-0.4

0.8

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

1990-19941995-19992000-20042005-20092010-2014

0.50.3

-0.5-1.3

-0.05

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

1990-19941995-19992000-20042005-20092010-2014

United States Japan Germany

France United

Kingdom

Italy

A key measure of how well an economy uses its existing capital and people is “total factor productivity” (TFP).

Data show a weakening of TFP, especially in the past 10 years.


SO ARE WE EXPERIENCING THE END OF GROWTH AND

PROSPERITY?


AI CAN BE A POWERFUL REMEDY FOR SLOWING GROWTH

Source: Accenture analysisindicates the change in that factor.NB:Δ

AI has the potential to overcome the physical limitations of capital and labour and open up new sources of value

and growth

Capital Labour TFP

Growth

Capital Labour TFP AI

Traditional Growth

Model

Adapted Growth Model

Δ Δ

ΔΔ Δ


AI CAN BOOST NATIONAL ECONOMIC GROWTH VIA THREE

CHANNELS

AugmentationIntelligent

Automation

Innovation

Diffusion

1 2 3


AI PROMISES TO BOOST NATIONAL ECONOMIC GROWTH

2.6

2.12.5

1.7 1.61.4 1.4

1.7

0.8

1.6 1.7

1.0

4.6

4.13.9

3.63.2 3.0 3.0 2.9

2.7 2.7 2.5

1.8

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

Baseline AI steady state

Real gross value added (GVA)

(%, growth)

AI has the potential to double annual growth rates in the countries we analyzed in terms of gross value added

(a close approximation of GDP).

Source: Accenture and Frontier Economics

19

CONCERNS AND FEARS AROUND AI

ARE ON THE RISE


AI IS ON THE CUSP OF TRANSFORMING EVERY ASPECT OF

HOW WE DO OUR JOBS AND LIVE OUR LIVES

HOW WE LIVEHOW WE WORK

• Labour market

disruption

• Future of jobs

• Erosion of individual

identity

• …

• Data privacy

• Bias and fairness

• Growing inequality

• Post-work society

• …


CONCERNS THAT TECHNOLOGY WILL DESTROY OUR JOBS

ARE NOT NEW

2011

David Ricardo,

The Principles of

Political Economy

and Taxation“…the substitution of machinery

for human labour, is often very

injurious to the interests of the

class of labourers.”

Karl Marx, Capital

“…this effect of machinery…

represented to be a

compensation to the working

class, is on the contrary a most

frightful scourge.”

John M. Keynes,

Economic Possibilities

for our Grandchildren“We are being afflicted with a new disease...

namely, technological unemployment.

This means unemployment due to our

discovery of means of economising the use

of labour outrunning the pace at which we

can find new uses for labour.”

1821 1887 1930 1952 20131989

Kurt Vonnegut,

Player PianoA novel that depicts a

world dominated by a

supercomputer and run

completely by machines

Paul Samuelson,

Ricardo was Right!Frey & Osborne,

The Future of

Employment:

How Susceptible

are Jobs to

Computerisation?

McAfee &

Brynjolfsson,

Race against

the Machine


HISTORICALLY, TECHNOLOGY HAS AFFECTED LABOUR

MARKET TRANSITIONS IN FOUR DIFFERENT WAYS

Substitution

EffectComplementary

Effect

Income

Effect

Price

Effect

Industrial robots took over

manual tasks such as welding,

painting and assembling

Computers improved efficiency

and created new jobs such as

computer-aided designers

Full-time earnings for British

workers more than doubled in

the seventy years after 1780

Transportation innovations and

development sharply pushed

down the cost of travelling


MH

NEW EMPLOYEE ROLES AND NOVEL TYPES OF WORKING

RELATIONSHIPS ARE REQUIRED

Human-only

activity

Machine-only activity

Humans complement

machinesAI gives

humans

superpower

Human and machine

hybrid activities

Le

ad

Em

pa

thiz

e

Cre

ate

Ju

dg

e

Tra

in

Su

sta

in

Exp

lain

Am

plif

y

Inte

ract

Em

bo

dy

Tra

nsa

ct

Ite

rate

Pre

dic

t

Ad

ap

t

The missing middle: new jobs, roles, and skills will be required as organizations use AI to augment and

transform the way they work

Source: Human + Machine: Reimagining Work in the Age of AI, Daugherty & Wilson, Harvard Business Review Press, 2018


AI BECOMES DEEPLY INGRAINED IN OUR EVERYDAY LIVES,

YET REAL CHALLENGES REMAIN REGARDING PRIVACY,

TRUST AND EXISTING BIASES

Source: Accenture Dynamic Consumers, 2017

Consumers want to be more engaged in managing

their data

Slaves to the algorithms?

of consumers believe it is important

to be able to review and control their

personal data online. 87%

3/4however, do not find it easy to

manage that data.

• Reinforcing existing stereotypes, social and

cultural segregation and exclusion

• Personalized recommendation leading to a

‘filter bubble”?


BUT AI COULD BE THE PIVOTAL TECHNOLOGY IN

ADDRESSING THE GRAND CHALLENGES OF HUMANITY

Inclusive education and personalized

learning

Empowering the disabled

Health management and disease

diagnosis

Fight against world hunger

Fight against crime

Urban traffic managementThe AI traffic system in Pittsburgh- Surtrac which can

detect traffic and allow traffic lights to adapt to real-

time conditions has helped cut travel time by 25%

Knewton, an AI-enabled adaptive learning platform leverages machine learning to personalize content to each user to improve academic outcomes.

Seeing AI, a free mobile application designed to support people with visual impairments by narrating the world around them.

DreaMed Diabetes, an Israel-based start-up, developed an algorithm which continuously monitors patients’ glucose levels and defines precisely when and how to adjust insulin levels.

Stanford University has instituted an AI program that uses machine learning to predict crop yield. The goal is to provide planters with information allowing them to make more thorough planting and food reservation decisions.

The Murder Accountability Project (MAP) has developed an online serial-detection algorithm. Established in 2015, this serial-detection system is designed to identify areas with significantly low clearance rates through the recognition of the homicidal patterns of individual serial killers.

26

How to realize AI’s promise while

controlling the peril

It all depends on how we manage

the transition to an era of AI


FORGING A DIGITAL NEW DEAL THAT ALLIES ECONOMIC

PROSPERITY WITH SOCIAL PROGRESS

Copyright © 2018 Accenture. All rights reserved.

• Imagine the future

• Address grand challenges

• Catalytic investment

• Regulatory innovation

• Reskilling for Jobs of the

Future

• Social and Technical skills

• Transition paths for affected

workers

• Life-long education system

• Responsible AI

• Missing Middle skills

• Regional & sectoral

adjustment

• Grassroots support

A digital 'New Deal’ - Three priorities

Big

Ambition

Big

Inclusion

Big

Skills Agenda

28

ANNEX


READY FOR HUMAN-MACHINE COLLABORATION ?

It will be important/very important to learn new

skills to work with intelligent technologies in the

next 3 to 5 years.


Employers underestimate the willingness of employees to acquire the relevant skills

62%

I believe intelligent technologies will create

opportunities for my work.

67%

ONLY 3%of executives say they intend to significantly increase

investment in training and reskilling programs in the

next three years.

Sample: More than 1,200 CXOs, more than 14,000 workers, 12 industriesSource: Accenture, Reworking the revolution, 2017


AI WILL BE BOTH A DISRUPTING AND ENABLING FORCE IN

THE WORKPLACE

Over the years, technology has made some jobs redundant,

but also led to the creation of new jobs and occupations:

Source: OECD, OECD Employment Outlook 2017

High-skill

occupations

Middle-skill

occupations

Low-skill

occupations

7.6

-9.5

1.9


Change of share of employment in the OECD area, 1995-

2015 (percentage points)• Free us up from 3 “D’s” – Dull, Dirty, Dangerous

But the impact of AI is more than automation; It’s

more about workplace augmentation

Legal analysis

AI surgical robotsVirtual assistants

• Make us more productive and innovative


AI RISKS EXACERBATING INCOME INEQUALITY AND

POLARIZATION OF SOCIETIES ACROSS THE WORLD

Technological progress has contributed the most to

widening income inequality across the globe from mid-

1980s to mid 2000s

(average annual percent change)

Source: IMF

-0.25 0 0.25 0.5

Change in inequality

Contribution of trade andfinancial globalization

Contribution of technology

Contribution of other

…And a new digital divide is looming on the

horizon

0

20

40

60

80

100

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

High Income Countries Middle Income Countries

Low Income Countries World

Internet penetration (%), 2000-2015

Source: OECD


LOOKING AHEAD, AI WILL POSE NEW CHALLENGES TO BOTH

REGULATORS AND BUSINESSES

Limits of Taylorism

Universal basic incomes

Robot taxes

Digital inclusivenessLearn with the machine

Transition paths to

new jobs

AI ECONOMIC SALVATION OR SOCIETAL PERIL? Talk_April 2018.pdf · Accenture Research Bath Royal...

Documents

Transcript of AI ECONOMIC SALVATION OR SOCIETAL PERIL? Talk_April 2018.pdf · Accenture Research Bath Royal...